Compact Genetic Algorithm

Compact Genetic Algorithms (cGAs) are a class of low-memory evolutionary algorithms that optimize by iteratively updating a probability distribution over potential solutions, rather than maintaining a large population. Current research focuses on analyzing their runtime performance on various problem landscapes, including extensions to multi-valued variables and investigations into the impact of genetic drift and algorithm parameters like step size and population size. Understanding the strengths and limitations of cGAs, particularly on challenging problems like those with "cliffs" in the fitness landscape, is crucial for improving their efficiency and expanding their applicability in diverse optimization tasks.